cockle doodle doo

I probably shouldn’t admit this, but my Day of Digital Humanities begins pretty much when I wake up. I sleepily reach for my iWhatever and check Twitter and email (usually in that order). Given everyone I follow is vaguely related to digital humanities (but then again, metaphysically speaking, who isn’t?), that first act tends to start the thaw of my lethargic and reluctant intellectual juices. Then I usually fall out of bed, stumble down the stairs, prepare milk for my daughters, light the gas fireplace (if applicable), and throw the switch on the espresso machine. That last gesture is particularly significant as the rest of my morning home routine culminates in a shot of viscous nectar of the goats (as pictured). Between the initial flicking of the switch and the eventual sipping of the ambrosia, I alternate between yawning, preparing myself, helping the girls get ready, meeting the needs of our dog, and musing about things DH (if you’ll allow me to alternate between more than two things, that is).

My Dog Joe

Actually, my dog’s name is Sablé – My Dog Joe is the name of the café I sometimes go to before heading to my office. It’s an ideal café in many respects: funky, homey feel to it, choice of good coffees that can match your ideological leanings, independent, lots of sunshine streaming through the window (if applicable – and it’s most certainly applicable today), etc. The only unideal thing is that it’s not within walking distance from home (but perhaps my monthly coffee budget is more reasonable because of this).

This morning I had several items on my todo list while I sipped coffee, including:

All of this was before Cyril Briquet arrived to discuss our work on Trombone (the underlying analytic module used by Voyeur). We tend to meet once or twice a week at My Dog Joe (often enough that we abbreviate it to MDJ). Cyril is a postdoc (funded primarily by SharcNet) working with me on making Trombone run 1) on HPC; 2) in the cloud. Initially we will work with Hadoop, but possibly also Cyril’s own Canopeer project. Since his arrival Cyril has done an amazing amount of work refactoring the code I’d written to improve the general architecture and to decouple the data structures from the storage strategies; this is a key step for allowing distributed computing and storage of a stateful(-like) web application.

At the moment Cyril is focusing on the tools package (the tools operate on data structures that have been prepared and stored elsewhere). The tools create data objects for results and then we (currently) use XStream to serialize the data objects as JSON or XML (that can be used by various clients, including Voyeur). This morning we discussed the interaction between tools in a distributed environment as well as API ideas for accessing tokens and types of various forms (case harmonized, stemmed, lemmatized, semantic, etc.).

Teaching Programming

I consider myself extremely fortunate to be able to explicitly teach technology with a humanities perspective in McMaster’s undergraduate Multimedia programme. I have doubts that I’d be anywhere as happy professionally if I had a more conventional position teaching, say, French literature (which corresponds to my academic background). There are some things I truly miss about a primarily literary perspective, but I usually find a way to impose some of that on our students anyway. I should be clear: it’s a bit of stretch to incorporate my core research interests (on text analysis and visualization interfaces) into much of our curriculum, but I still think of myself as teaching digital humanities.

I just finished a tutorial for my Programming Fundamentals course. The course is structured with 2 hours in a lecture classroom and 1 hour of tutorial, where students work on weekly self-guided “checkpoints” (or exercises). I’ve made tutorials optional – if you can do the checkpoint, then please, by all means, enjoy the sunshine; if not, I’ll be in the lab with other students to try to help you.

I find it a big challenge to teach programming to our students, in part because the syntactic requirements are usually so rigid (and they’re not used to that), in part because I choose to try to teach programming concepts rather than a more pragmatic approach of just coding something. Quite frankly, I find that most students want to be able to program, but they’re not always willing to invest the time an energy required to work within this new paradigm – some also simply lack the confidence to do so (confidence is important in language acquisition in general). At the same time, I get very excited when I see students doing innovative work and when I see them feel empowered.

Next up this afternoon is a set of meetings with our senior thesis students. This is essentially an independent study course the provides a capstone experience for their Multimedia studies. The work individually or in pairs and prepare a major project. There are no classes in this course, students just setup meetings with their supervisor. For whatever reason there are a lot of food-related projects this year. There’s also a drum-kit rhythm game essentially built from scratch (the hardware part), a cool iPhone application, and a custom-built MMORPG, among others. Damn I love my job.

Doing Text Analysis

I first got into text analysis in the early 1990’s. At first I had very little relevant programming skills, so my range of activities was limited to using existing tools (like TACT), and reading about what others had done (Paul Fortier was a strong early influence). The focus at that time was on texts and their study. During my graduate studies I became increasingly involved in the development of tools, culminating in my PhD dissertation, which was a hybrid work on 1) Georges Perec’s La Disparition and 2) tools that I’d developed to help me study Oulipian texts. During the course of subsequent years, including my first job at the University of Alberta and my current job at McMaster University, the balance of my work has tilted toward the development of tools to the detriment of actually using them to study texts (one notable and very rewarding exception is work with Geoffrey Rockwell on In other words, I started with a mostly literary purpose and I’ve slowly migrated toward a mostly tool development focus. Truth be told, I actually like to lead a more balanced professional life, but I find it very difficult to find the time to nourish and sustain both perspectives.

That’s nice, Sinclair, but what’s it got to do with Day of Digital Humanities? I’m glad you asked. I’m currently working with Stephanie Posthumus (a scholar of twentieth and twenty-first century French literature and, incidentally, my spouse), on an analysis of two works by the French writer Michel Houellebecq: Les Particules élémentaires and La Possibilité d’une île (which we had scanned for personal research purposes by McMaster Library’s robotic scanner). We’re especially interested in studying the interplay between nature and technology in the two works.

I must say that every time I embark on a real literary project I’m humbled by the primitiveness of our text analysis tools (or I guess I could say that I’m awestruck by the endless idiosyncrasies of literary texts). Our tools do a number of things quite well, but I almost always need them to do something slightly (or completely) different. This actually provides me with a healthy reminder of how much work remains to be done (and perhaps why our recalcitrant non DH colleagues aren’t so wrong). I’m partly saying that every situation of literary interpretation and analysis might require a somewhat unique tool that would be impossible to anticipate in advance, but I’m also saying that we have a great deal of work still to do in providing general purpose tools, especially ones that have any semantic functionality built-in. Many days of DH ahead.

Wrapping Up

Well I started blogging today with a dirty little confession about how early by DH day starts, and I have a similar confession about how late it lasts: once the kids are in bed, I usually go back and do some work. I go through periods where I read fiction in the evening, either because I have a lighter work load or because I’m just more willing to ignore my work – but alas, not today. One of the last things I wanted to do was to create a Voyeur corpus of the DayOfDH, where each document represents an author. The other thing I wanted to do was to create a corpus based on all the posts, but that may have to wait, as I’m not sure how to grab all the posts in chronological order (with a minimum of work).

Once again I enjoyed the DayOfDH immensely, not only as a chance to see what other are doing, but also as a chance to reflect more deeply on what I’m doing (I don’t blog much anymore, that itch has been mostly replaced by Twitter, though of course Twitter is a different kind of scratching). Many thanks to the folks at Alberta for all their hard work in making this happen!